The Shapiro-Keyser Cyclone Model


Appearance in METEOSAT Second Generation imagery

The Shapiro-Keyser cyclone model has been developed in the late 80ies, benefitting from data made available by meteorological satellites and numerical model simulations. This cyclone model has been named after the authors of the study that first presented this conceptual model (M. A. Shapiro and D. Keyser, 1990). The authors discovered that for some types of cyclogenesis, the meteorological data seemed not to confirm the classical Norwegian cyclone model in which the occlusion process results from a merging of the cold and the warm front. Instead, they observed a cold frontal fracture at the place where the cold front should smoothly turn into the warm front (see Fig. 2b).

Figure 1: Rapid cyclogenesis developing over the northern Atlantic (8 - 9 January 2018). The pressure in the center of the low drops from 961 hPa to 946 hPa in less than 24 hours.

In the above Airmass RGB loop (Fig. 1), we can observe the formation of a cloud spiral (commonly named occlusion) which is not the result of a classical occlusion process as described by the Norwegian cyclone theory. The formation of this "occlusion" cloud band is best described with help of the conveyor belt theory (see the next chapter).

The Shapiro-Keyser cyclone model in a way complements the classical Norwegian cyclone model. It is often observed at cyclones forming over the oceans during cases of Rapid Cyclogenesis.

Figure 2: Comparison of the Norwegian model with the Shapiro-Keyser model of cyclogenesis. Both schematics depict lower-tropospheric geopotential height (e.g. 850 hPa) and potential temperature.

The Norwegian cyclone model (a) shows the development of a cyclone starting from a disturbed frontal zone (stage I) in which an amplifying wave [ (II) and (III)] forms a cold and a warm front. The faster moving cold front catches up the slower moving warm front, both fronts form an acute angle. Finally, both merged fronts, the so called occlusion, wrap around the deepening low pressure core.

The Shapiro-Keyser cyclone model (b) starts from the same disturbed frontal zone (stage I). But instead of catching-up the warm front, the cold front moves perpendicular to the warm front (a.k.a. T-bone structure) but without even touching it. The gap in between the cold and the warm front is filled by warmer air masses from the warm sector penetrating the low pressure core and forming a warm seclusion. While for the classical Norwegian model, the merging of the cold with the warm front forms the occlusion, in the Shapiro-Keyser cyclone model, the same cyclonically curved cloud band is depicted as a warm front forming the so called "hammerhead" (see Fig. 3). An occlusion per se does not exist in the Shapiro-Keyser model. The emerging cloud band (S-K refer to it as "bent-back warm front") starts to rotate around the deepening low pressure core (stage III and IV). In the final stage (IV), cold air from behind the cold front starts to wrap around the warm core and the cold front attaches to the warm front (i.e. the frontal fracture closes).

From a satellite point of view, the final stage, i.e. the cloud spiral wrapped around the low pressure area, has a very similar appearance in both conceptual models. The main difference lays in the earlier stages when the low pressure system develops.

Analysis of a typical Shapiro-Keyser cyclone by means of the conveyor belt theory

Conveyor belts offer a slightly different view on the dynamics of cyclogenesis. They show the transport of air masses relative to the displacement of the frontal system, that means that we see the air streams as an observer moving with the system would see them. 3 different conveyor belts can be diagnosed:

  1. The warm conveyor belt which usually moves northward ahead of the cold front and then turns eastward, gliding up the warm front. This conveyor belt usually rises along it trajectory.
  2. The cold conveyor belt, which has its origin in the cold air beneath the warm front. The air moving along this conveyor belt is also directed upwards.
  3. And finally the dry intrusion which consists of sinking dry air behind the cold front.

Figure 3: Schematic showing the position of the conveyor belts involved in the cyclogenesis process. IR10.8 μm image from 9 January 2018 at 06:00 UTC.

The warm conveyor belt reaches levels of 500 hPa and higher when it turns east up onto the slope of the warm front. It moves along the lines of the geopotential height and hence follows the high pressure ridge east of the trough.

In case of a Shapiro-Keyser cyclone, the dry intrusion moving northward at the rear of the cold front penetrates the warm core of the cyclone and starts to wrap around it. Because the dry intrusion is still warmer than the cold conveyor belt, the warm front character of the bent-back warm front is preserved. We find warm air advection all along the inclined slope of the warm front (Fig. 4, top right) and a clear temperature gradient (Fig. 4, bottom right).

Figure 4: Vertical cross section through the cold conveyor belt (green line). Top right: Temperature advection; bottom right: temperature. Date: 9 January 2018, 06:00 UTC.

When the low pressure system deepens and starts to show closed contour lines of the geopotential height at higher levels (> 500 hPa), the warm conveyor belt splits and a new branch develops (see Fig. 5). The splitting of the warm conveyor belt can be observed in the IR imagery. The cold conveyor belt protruding from beneath the easterly branch of the warm conveyor belt and wrapping around the low pressure area is much lower than the warm conveyor belt at high altitudes. When the warm conveyor belt splits, it covers the western part of the cold conveyor belt and both, the cyclonic cloud spiral and the warm front are depicted as a uniformly high cloud shield in the IR10.8 μm imagery.

Figure 5: Schematic showing the splitting of the warm conveyor belt covering the cold conveyor belt at the mature stage of cyclogenesis. IR10.8 μm image from 9 January 2018 at 12:00 UTC.

The warm core seclusion of the Shapiro-Keyser cyclone

The process of the warm core seclusion during the cyclogenesis process in Shapiro-Keyser cyclones is tightly connected to the cold frontal fracture. Figure 6 shows the seclusion process which leads to the warm core of the cyclone. During the encapsulation process, the thermal gradient, characteristic for a cold front, disappears at the northern tip of the cold front before reaching the warm front. Instead, we find the highest temperature gradients where the cold conveyor belt wraps around the low center.

Figure 6: Isolines of the temperature at 850 hPa and the position of the cold and warm front in a Shapiro-Keyser cyclone. SEVIRI IR10.8 μm images from 8 and 9 January 2018 (18, 00, 06 and 12 UTC) from top left to bottom right.

Towards the end of the seclusion process, the northern part of the cold front rejoins the warm front (T-bone structure). The sinking cold air masses behind the cold front (dry intrusion), begin to wrap around the warm core. From the satellite point of view, the final stage of the Shapiro-Keyser cyclone is similar to the classical Norwegian cyclogenesis.

The Large Scale Flow

Schultz & Wernli state that the nature of the large-scale flow (i.e. mid/upper tropospheric flow) is an important factor in determining the cyclone type. Cyclones embedded within a diffluent flow (i.e. jet-exit region) tend to evolve like the Norwegian cyclone model, whereas cyclones embedded within a confluent flow (i.e. jet-entrance region) tend to evolve like the Shapiro-Keyser cyclone model.

The example from 8 and 9 January 2018 shows the formation of a Shapiro-Keyser cyclone in the entrance region of a jet streak (see Fig. 7).

Figure 7: SEVIRI IR10.8 μm image from 8 January 2018, 12:00 UTC (left), 18:00 UTC (middle) and 9 January 2018, 00:00 UTC (right). The red lines depict the jet axis and the green stars the core of the low pressure at sea surface level.

  1. At 12:00 UTC, the surface low pressure center (green star) of the Shapiro-Keyser cyclone is located in the entrance region of a jet streak (red line between the green star and Iceland).
  2. 6 hours later (18:00 UTC), the developing low is still in the jet entrance region. The jet streak starts to extend southwards (red dotted line).
  3. Again 6 hours later (00:00 UTC), the Rapid Cyclogenesis shows 2 separated jet streaks (red lines).

Because the large-scale flow is not constant in time, cyclones can change from one type to another. For example, many initially Shapiro-Keyser-like cyclones may develop occluded fronts late in their life cycles, becoming more Norwegian-like, as the initially confluent flow becomes more diffluent during cyclogenesis (Schultz and Wernli, 2001).

Key Parameters

Regarding the key parameters of the Shapiro-Keyser cyclone, please refer to the chapter of the Rapid Cyclogenesis. While the key parameters are showing the same pattern as for the classical Rapid Cyclogenesis, the main difference concerns the large scale flow in the mid- and upper troposphere as mentioned in the previous chapter.

The sting jet

Sting jets are areas of high wind speed maxima descending from about 700 hPa to the surface level, located equatorward of the low center near the tip of the cloud head at the end of the bent-back warm front. Sting jet are not exclusively observed with Shapiro-Keyser cyclones, but they are frequently linked to the latter. One hypothesis for the generation of sting jets is that when the cold conveyor belt, which wraps around the warm core of the cyclone, merges with the dry intrusion (see Fig. 8), cloud droplets from the southern tip of the bent-back warm front evaporate in the dry air, and in doing so, cool the ambient air. This cooling process leads to destabilization and results in subsidence of the air flow within the merging zone; the so called sting jet.

An acceleration of the downdrafts involved in the sting jet can be caused by the increasing horizontal pressure gradients when moving down to lower levels. The downdrafts can reach the ground with very high wind speeds which are among the strongest in the whole cyclonic system.

Figure 8: Schematic showing the merging zone of the bent-back warm front with the dry intrusion. SEVIRI IR10.8 μm image from 9 January 2018, 00:00 UTC.

Sting jets are difficult to predict with numerical models because of their limited spatial extension and the way each individual low-pressure system develops. A shift of 50 km will menace a region which previously had been outside the danger zone.

At 00:00 UTC on 9 January 2018, model fields give some hints that a sting jet occurs at that moment. Figure 9 (top images) shows a region of high wind speed at 925 hPa and at ground level at the southern tip of the bent-back warm front. In the vertical cross section (see Fig. 9, bottom right), the highest wind speeds can be found in the layer between 800 and 925 hPa. The IR10.8 μm image indicates a small clearance of the cloud cover at the indicated wind speed maximum. This can be an additional evidence for the presence of a sting jet.

The vertical cross section on the left hand side of Figure 9 shows descending vertical motion (dotted blue lines) in the area of the wind speed maximum. Near ground level, wind speed decreases probably due to surface friction.

Figure 9: Top images: Wind barbs from ECMWF model at 925 hPa and at ground level, the yellow lines depict the 30m/s isotach; SEVIRI IR10.8 μm image from 9 January 2018, 00:00 UTC.
Bottom images: ECMWF vertical cross sections along the green arrow.


Martinez-Alvarado, O.: Sting Jets and the Diagnosis of Conditional Symetric Instability, 13th AMS Conference on Mesoscale Processes, August 2009, Salt Lake City.

Schultz, D. M., and Browning, K. A.: What is a sting jet?, Royal Meteorological Society, Weather, March 2017, Vol. 72, No. 3.

Schultz, D. M., D. Keyser, and L. F. Bosart, 1998: The effect of large-scale flow on low-level frontal structure and evolution in midlatitude cyclones. Monthly Weather Review, 126, 1767-1791.

Schultz, D. M. and H. Wernli, 2001: Determining Midlatitude Cyclone Structure and Evolution from the Upper-Level-Flow

Shapiro, M. A., and D. Keyser, 1990: Fronts, jet streams and the tropopause. Extratropical Cyclones, The Erik Palmén Memorial Volume, C. W. Newton and E. O. Holopainen, Eds., Amer. Meteor. Soc., 167-191.

Young, M. V., 1995: Types of cyclogenesis. Images in Weather Forecasting, M. J. Bader, G. S. Forbes, J. R. Grant, R. B. E. Lilley, and A. J. Waters, Eds., Cambridge University Press, 213-286.